What is expert system in computer science

Expert system

info

An expert system (XPS) is a computer-aided planning system, i.e. a software system that has the knowledge of experts in a certain (delimited) problem area (knowledge domain) and is able to apply this knowledge to solve problems. It is also known as a knowledge-based system and represents an important area of ​​research on artificial intelligence.

Areas of application

Expert systems can be used in different areas. So they are suitable for:

  • the interpretation (interpretation systems), i.e. the evaluation or analysis of information, e.g. operational data (data analysis), acoustic (speech processing) and optical information (image processing);
  • the diagnosis of system states (diagnostic systems), i.e. the determination of error states and their causes in technical systems (e.g. data processing systems and networks, vehicle engines) and in biological systems (e.g. medical diagnosis);
  • the construction of certain objects according to given specifications (construction systems), e.g. draft (design) of a construction plan with the help of CAD systems;
  • the planning of action sequences, e.g. when using robots, in production control or in business planning (planning systems in sales, in procurement and warehousing, in production and logistics);
  • advice (advisory systems), e.g. in business planning (e.g. investments, investment decisions) and
  • tutoring (tutoring systems), i.e. systems of teaching and learning (e-learning systems).

properties

  • Expert systems are computer-aided planning, analysis and construction systems (problem-solving systems);
  • are interactive program systems;
  • relate to a delimited sub-area of ​​human knowledge, e.g. in the operational area in production or in logistics;
  • use not only factual and rule-based knowledge, but also heuristics and vague knowledge to solve problems;
  • derive problem solutions from conclusions from stored knowledge (logical information processing);
  • are particularly suitable for working on "diffuse" areas of application that are characterized by fuzzy and incomplete knowledge (unstructured problem areas);
  • simulate the work of human experts;
  • guide the user through "natural language" dialogue;
  • explain their approach and the derived results;
  • show "intelligent" behavior in problem solving and
  • are capable of learning.

System architecture

An expert system is not a monolithic unit, but can be explained by functional components. In addition to the two core components, the knowledge base (contains the knowledge of the problem area) and the inference component (problem-solving component for the logical derivation of solutions), it has a dialogue and explanatory component. A knowledge acquisition component is required to set up and update the knowledge base. It is also desirable to have a learning component that can understand the ability to learn and thus automatically expand and improve the knowledge base. In addition to the necessary integration into the data processing system, interfaces to application systems such as B. to database systems and also to other expert systems, makes sense.

Performance potential

The performance potential of the expert systems lies primarily in problem solving on the one hand and in the ability to dialogue and explain on the other. A wide variety of forms of knowledge representation can be used to solve problems, such as rules, frames and objects. Inference processes and search strategies are used to find solutions.

Areas of application

In addition to the use of expert systems in operational application areas, e.g. in procurement, production, sales and human resources, the systems are of great strategic importance. Expert systems or knowledge-based systems are used to systematically process and secure the existing knowledge of a company (knowledge retention), to distribute it (knowledge multiplication) and to make it available at the right place at the right time. The systems are therefore important and powerful tools for operational knowledge management.

Literature:

Gabriel, Roland: Knowledge-based systems in operational practice. London: McGrawHill, 1992.

Harmon, Paul; King, David: Expert Systems in Practice. Perspectives, tools, experiences. 3. Edition. Munich: Oldenbourg, 1989.

Mertens, Peter; Borkowski, Volker; Geis, Wolfgang: Operational expert system applications. 3. Edition. Berlin et al .: Springer, 1993.

author


 

Prof. Dr. Roland Gabriel, Ruhr University Bochum, Chair for Information Systems, Universitätsstrasse 150, 44780 Bochum

Author info


Item Actions